At the Joint Statistical Meetings in Boston in early August, Gareth Roberts will give the inaugural Blackwell Lecture, awarded by the Institute of Mathematical Statistics. The lecture has been introduced in honour of the extraordinary career in statistics and probability of Professor David Blackwell, who died in 2010 at the age of 91. The lecture should present material with some connection to David Blackwell's work, although since Blackwell's work was so broad, this turns out not to be very restrictive. Gareth's lecture will talk about work on Monte Carlo methods for stochastic processes, which has close connections to one of Blackwell's most famous theorems: The Rao-Blackwell Theorem.

It is a tremendous honour for Gareth to be chosen as the first-ever Blackwell Lecturer. The title and abstract of his lecture are given below.

Rao-Blackwellization for Improved Monte Carlo for Stochastic Processes

David Blackwell's celebrated theorem with Rao compares the variance of an estimator X with that of its conditional expectation given a further variable S; that is, Y = E(X|S). Blackwell's motivation was to demonstrate that the best estimators should be functions of the simplest sufficient statistic we can find. However the Rao-Blackwell theorem has also found extensive use in computational statistics by providing a powerful and general recipe for improving on Monte-Carlo estimators by so-called Rao-Blackwellisation. This talk will discuss widely adopted, simple Rao-Blackwellisations for Markov chain Monte Carlo and Sequential Monte Carlo samplers, as well as more complex random Rao-Blackwellisations for estimation of expectations of functionals of stochastic processes (particularly univariate and multivariate diffusion processes).

The lecture will draw on joint research with Alex Beskos, Paul Fearnhead, Krys Latuszynski, Omiros Papaspiliopoulos and Giorgos Sermaidis.